Taming the Most Detailed Product Data Imaginable

Some things are a given today. One is computing is cheap and powerful, and it is getting cheaper and more powerful. Another is the dropping price of industrial sensors. Combine this with easier ways of moving around data from those sensors and you get lots of data: Terabytes of data.

These trends are leading to one of the more interesting problems of the 21st century: How to make sense of it all.
How can makers of complex products—airplanes, passenger cars and toaster ovens—use vast streams of bits and bytes to make them better?

Enter the Digital Twin, fed by its close relative the Digital Thread. They are starting to catch on.

“Demand for creating Digital Twins was almost nothing three or four years ago,” said Doug Macdonald, product marketing director for Aras Corp. “Today a significant segment of our customer base is building or using them.”

Aras is following a unique business model with its open-source Innovator, a solution for product lifecycle management (PLM). The open source PLM model is one Macdonald thinks is a critical enabler
for Digital Twins. Often, Aras’ customers lead the company in creating new applications with the Aras Innovator.
“Digital Twin is a good example of where [our] customers created something we first see at one of our events,” he said. “There is a lot of creativity going into making Digital Twins.”

So, what are Digital Twins?

There are many definitions of the concept being kicked around, Macdonald said. Vendors are talking about them in different ways. Their descriptions are often tailored to the markets they serve, playing to the strength of the tools they offer.

“Our view of a Digital Twin is that it defines the actual configuration of the actual asset—an aircraft, an engine or something equally complex,” Macdonald said. “It is much more than the CAD model. Generalized CAD is only something you plan to build, not a model of the actual thing.”

The likelihood that it represents what is actually in the marketplace—as-built versus as-designed—is non-existent. “You also need more information, right down to the exact serial number of each part in each shipped model,” he said.
Adding data shipped via the Internet of Things (IoT) from sensors tied to the product to the Digital Twin now requires context. “If you have IoT data streaming off the product, unless you know the configuration you don’t have a context,” Macdonald said.

Take the example of an aircraft engine. If you have a sensor on the fuel pump and the outlet pressure is high on a particular batch of engines, that may or may not be normal. It could be that a batch of engines were modified such that high outlet pressure is now normal, and if the Digital Twin model captured that, the operator would be fully informed.

Design and visualization

“If you only have a product definition in digital form, then you have a sibling—you don’t have a twin,” said Mark Taber, VP of marketing for PTC. “We view the Digital Twin as all of the product data that you would manage throughout its lifecycle—the complete experience of the product that includes its time in the factory and in the field.”
PTC has in recent years embraced IoT has one of its core technology offerings with its ThingWorx platform.

“PTC has invested about $1 billion in acquisitions and organic R&D in the last three or four years,” said Mark Gallant, senior director of marketing for IoT manufacturing solutions at PTC.

That includes investing in Digital Twins and the Digital Thread.

Organizing how to transmit data and associate it with the Twin is a vital role of the Thread. “It is not enough to put data into an unorganized lake and expect to mine it for any sort of insights. It would not scale. It is not practical,” he said. “Context and associativity are critical, or you cannot make sense of all that information.”

This is why PTC offers technologies like ThingWorx Navigate. It gives product information access to not only engineering teams but also manufacturing, service and sales teams. PTC is now categorizing each team’s most pressing product questions by task and building out-of-the box Digital Twin apps to get answers to those questions.
“On the Creo side, [our CAD origination product], the design process begins with these stakeholders defining the ‘whole product,’ its subsystems, interfaces and the desired data streams,” Taber said.

Engineering sits side-by-side with marketing, service, packaging, operations, sales and finance deciding the IoT strategy and the customized data streams needed to support that strategy.

PTC is also seeing a huge pull from its customers in new Digital Twin and Digital Thread projects. “We are seeing huge demand within the manufacturing community,” Gallant said.

One of the reasons is the changing workforce.

Professionals now hitting the stride of their careers grew up with advanced and portable digital technology. They expect easily consumed information at their fingertips with tablets and smart phones. They expect easy ways to make sense of the torrent of data.

That is why PTC created ways to employ augmented reality (AR) and virtual reality (VR)—to make that data easy to understand. It helps put the data in context, an important but tricky endeavor.

“What we can enable now is using a PC to perform 2D and 3D design reviews and then use augmented reality to send it to the factory and into the field where it is being used,” he said.

Data, simulation and insights

In a series of moves that might be construed as building capability in creating Digital Twins and Digital Threads, Hexagon Manufacturing Intelligence (HMI) acquired a number of firms beyond its base in metrology equipment and software. Adding to its metrology, CAM programming and forming simulations base, the latest acquisition is MSC Software. MSC provides of a variety of CAE simulation tools, such as the ADAMS multibody package and MSC Nastran finite element software.

“The connection to the real world through our metrology offerings is why MSC made sense to us,” said Brian Shepherd, senior VP for software solutions at HMI and CTO for MSC Software.

Metrology equipment is moving to the shop floor and collecting data as complex products are made. The next step is to use that metrology data in Digital Twin projects.

“What MSC technology [adds to our portfolio is] material behavior, manufacturing process simulation and part and product performance simulations,” Shepherd said. CAE simulation data now becomes part of the comprehensive picture of a Digital Twin. Using metrology and simulation together, engineers can now predict performance of the as-manufactured product, as well as analyze trends.

“For example, in the automotive world, we have used physical data to correlate crash models to ensure accuracy, and then simulated new scenarios to ensure vehicles are safe under other conditions,” said Leo Kilfoy, GM, engineering lifecycle management, at MSC Software. Shepherd noted that the firm is incorporating other manufacturing simulations, such as forming from its FTI unit, as well.

One of the more challenging data management problems emerging with fast, cheap computing is the vast number of simulation runs design engineers can exercise.This led to the need for simulation data management, which SimManager from MSC performs.

But just as important as having simulated data is combining it with the measured data in a Digital Twin. Using measured data along with data from running “what if” simulations would lead to unique insights, Shepherd said.
“We are actively connecting the metrology world from data available in our, say, PC-DMIS metrology offering with CAM packages from Vero and now MSC via its SimManager,” he said.

“We think that to create a really useful Digital Twin, you need a couple of fundamental technologies. One is CAE to mimic and predict performance beyond measured data,” he added. “A second one is technology for data orchestration, moving data from real world into simulated data and between silos within organizations themselves.”
Manufacturers are looking to break down those silos that have crept up in their organizations, and the Digital Thread and Digital Twins are an ideal way to do that.

Siemens PLM has fully embraced both the concept and potential of Digital Twins and Digital Thread. It has a comprehensive set of tools ripe for implementing them, using real and simulated data. These include NX CAD, Teamcenter for data management, and a host of Simcenter applications for CAE simulation, as well as test data analysis and management tools.

The company makes the point that there are, in reality, three Digital Twins that manufacturers will find useful: Product, Production and Performance Twins, said Ravi Shankar, director of global simulation product marketing.
Simulation models help validate performance ahead of deployment. They can then validate and commission manufacturing processes ahead of production startup.

Simulations refine the Digital Twin as data is collected, analyzed and fed back into them. “We employ models built around single physics, multiphysics and co-simulation combined with testing to increase and improve realism in the models that comprise a Digital Twin,” Shankar said.

He also agrees that creating a Digital Thread is vital, and Siemens offers that capability, as one might expect from a PLM provider.

“You need to make sure throughout [the process] there is a thread for engineers to understand which parameters were decided at what time and what is affected by that decision,” he said.

A comprehensive twin tied to a Digital Thread will then provide the ability to explore and extract insights.
This is accomplished both from the results of CAE and collected data from in-field and as-manufactured performance. “We see realism, continuity, and exploration as being critical elements for our customers to mature their digital twin capabilities,” he said.

Digital Twins and new business models

Digital Twins and Digital threads are creating opportunities for new business models.

“Many people today are challenged with inventing a new business model that does not include selling the product, but the service that product provides,” said Tom Maurer senior director, strategy at Siemens PLM.

“Selling ‘X-as-a-service’ is the hot thing with our customers right now,” Maurer said.

The aircraft engine industry has embraced this by selling hours of engine uptime instead of engines.

One company with which Maurer is familiar sells “flow-as-a-service,” using IoT-enabled valves, he said. Another company he follows makes industrial cranes, and it is exploring how to sell “lifting-as-a-service.”

Selling “X-as-a-service” is enabled or facilitated by Digital Twin and Digital Thread. They provide a means for a company to track performance and understand costs and profitability.

Digital Twin technology is in its infancy. Its attractions are obvious. It can make sense of large data sets, whether generated from tests, field data or CAE simulation—or all combined. It can break down communication barriers by providing a single, comprehensible focal point for all.

It will foster innovation both technically and for new business models.

“Companies that are laggards in developing Digital Twins risk being left behind,” Shankar warned.